Sensitivity of Extreme Hurricane Winds to Climate Change

点击翻译按钮获取中文摘要

Advances have been made in modeling extreme hurricane winds regionally. For example, the return period of a Katrina-like storm is 14 years along the entire U.S. coast as estimated from an extreme-value model derived from reliable landfall reports. But what are the return periods of hurricane winds at specific locations, like New York City? This question is more difficult to answer, because of the rarity of historical storms with sufficient intensity affecting this location. In this project extreme hurricane winds will be modeled locally based on new insights into the scaling behavior of the parameters of the extreme value distribution. This technology will allow users to condition wind exceedance probabilities on climate variables, such as ocean temperatures and steering currents, in order to quantitatively assess which cities are most sensitive to climate variations, in terms of their risk from hurricanes.The goal is to better understand how and to what extent local hurricane risk is affected by climate. The objectives are to develop and implement the technologies for anticipating extreme winds along the coast of the United States. The technical problems to be solved include (1) determining the proper model for the available data, and (2) accounting for the variable levels of uncertainty in the data records. The scientific problem is to understand how sensitive local extreme hurricane activity is to climate.A systematic approach to data modeling will be taken and the models will be made available to the scientific and risk management communities.The broader impacts of this work are a better understanding of the hurricane threat to the United States and elsewhere, a new tool kit for data modeling in the climate sciences, and estimates of hurricane return periods for any intensity at any location. Students will be educated in the theory of climate and in the application of modern statistical methods for hurricane climate research. Open-source software will be used, and students will be trained in the areas of applied computing and statistics. This work will advance and promote a broadening of quantitative data analysis that encompasses data models and Bayesian methodologies. Results will help to improve the way statistics is integrated into research and education in the climate and related sciences.